Variability in Slab Structure and Behavior Within and Among the South American and Eastern Mediterranean Subduction Systems

Abstract

As subduction science advances, it is becoming increasingly clear that the conventional model of subduction driven by cold, strong, dense slabs sinking cohesively in the mantle is insufficient to explain observed heterogeneities across a range of scales. The breadth of observed features of subduction zones is expanding at the same time as along strike variability is revealed within individual subduction systems. Not surprisingly, the driving components of subduction zones, and by extension this variability, are the subducting slabs. Thus, in order to revisit the traditional model of subduction, this dissertation uses seismic imaging techniques in four distinct studies to characterize the structure and behavior of the subducting slabs across two extensive subduction systems: South America and the Eastern Mediterranean. These two systems serve as the ideal test cases for characterizing variability within and across subduction zones because they represent end-member styles of subduction: Andean-type subduction and Mediterranean-type subduction, respectively. The first study presents the results of a regional-scale teleseismic tomography inversion surrounding the Sierras Pampeanas of Argentina. This study targets the observable response of the slab-mantle interaction to subduction of the hot spot-generated Juan Fernández Ridge, identifying entrainment of hot subslab asthenosphere and induced tearing of the subducting Nazca slab. Through this study I also present a model for explaining the infrequent distribution of subslab slow seismic velocity anomalies in subduction systems globally. The second study presents the results of a continental-scale teleseismic tomography inversion across the Anatolian sub-continent in the Eastern Mediterranean. This study targets the response of the slab to the major tectonic transition seen in central Anatolia from ocean subduction in the west to continental collision in the east. The resulting seismic images show an increasingly deforming slab from west to east that illustrates the evolutionary process of subduction termination. The third study presents new software developed to automatically extract slab geometry from teleseismic tomography models. The efficacy of the software is presented through applications to the Juan de Fuca/Gorda and Nazca slabs, signifying its broad utility in interpreting tomographic images. The fourth study utilizes this new software along with a continental-scale teleseismic tomography model across South America to present a detailed model of Nazca slab geometry from the surface into the lower mantle. The results of this study reveal that the slab penetrates the 660 km discontinuity uninhibited while instead interacting with a lower mantle discontinuity. This result has far reaching implications for geodynamic models of South American evolution. Together, these studies reveal a broad consistency across subduction zones – that slabs do not remain cohesive during their descent in the mantle. Instead, the presence of heterogeneities in the downgoing plate induces slab tearing and possibly fragmenting within the upper mantle. Despite the relative complexity of the multi-plate Eastern Mediterranean system compared to the two-plate South American system, both systems display significant variability along the margins, including regions with varying degrees of slab tearing. This observation directly contradicts the traditional view of cohesive, strong slabs sinking in the mantle, which has related implications for the role of subducting slabs in mantle convection and material recycling over geological timescales.